An elementary electronics question.

[Robin GrahamParticipant @robingraham42208]

Posted by Michael Gilligan on 08/11/2022 11:38:32:

Looks like we need to see the string of lights that Robin is using

A picture is worth a thousand assumptions !

MichaelG.

I apologise for the inadequate diagram in my OP – what I represented as a single LED is in fact a string of 12 connected in parallel with a single limiting series resistor:

I can't see the colours on the resistor clearly enough to identify it from them, but measuring gives 3.1 ohms. Shorting the multimeter leads gives 0.3 ohms (1 sf) so I'm guessing the resistor is 2.7 ohms.

Anyhow, as I said earlier, from my POV my questions have been answered and my thanks again to contributors. With my new-found understanding I can now see better ways of creating the effect I want using a combination of 12V lighting and mechanical components and high intensity LEDs so I'll probably not use the fairy light string anyway. It's served its purpose as an example!

Jeff – thanks for the pointer to Forrest Mims, I'll have a look.

Robin.

[John OlsenParticipant @johnolsen79199]

LED tapes have got even more sophisticated than that. The ones I am playing with have 3 coloured LEDs, or sometimes three colours plus a white one, with a little chip for each group of 3 or 4 LED's. The chip controls the colour of its LEDs, and each chip has a data in and out connection, daisy chained along the whole tape, which can be many LEDs long. You squirt a long serial data steam in, the first chip reads and deletes the first instruction and then passes the rest on. The result is a strip of maybe 300 or so LEDs, each individually addressable for colour and brightness. This all happens at 100's of kHz, so you can get smooth animations of brightness and colour down the strip. All very sophisticated, and surprisingly cheap, although of course total overkill for the initial posters query.

John

[Michael GilliganParticipant @michaelgilligan61133]

Thanks, Robin

Exactly what I expected to see [but we do like ‘avoidance of doubt’]

MichaelG.

[Anonymous]

Posted by Nealeb on 08/11/2022 11:33:01:

…the incremental cost of adding SM resistors can't be that much…

It doubles the time needed on the pick and place machine, every fraction of a cent counts.

Andrew

[Michael GilliganParticipant @michaelgilligan61133]

Robin,

This is a slight digression … but I think it’s in the spirit of this discussion

Have a look at this little board: **LINK**

https://www.adafruit.com/product/757

I think you will agree that it does a useful thing, within a tidy form-factor, at a reasonable price.

MichaelG.

.

.

Now consider that I have just received ten of ostensibly the same board, shipped from China, for a total cost to me of £7.17

Edited By Michael Gilligan on 09/11/2022 12:18:58

[Andy AshParticipant @andyash24902]

I've not read the whole thread, so forgive me if I've repeated something someone else said. Looking at this, it's an ideal application for a "buck regulator". The problem you have is the relatively high power dissipation dropping from 12 to 2ish volts at 270mA.

That's always going to be 2.7 watts in any linear situation. Its not that you can't get a resistor or transistor to handle that, but it's wasteful – especially if you're using a 12V battery.

You can get a buck regulator ready made circuit from ebay, but you can make one too. Broadly it is an oscillator which switches a pass transistor through a diode and an inductor in a PWM fashion. You could use a 555 timer for the oscillator. There is a feedback loop and an error amplifier which are used to control the mark space ratio of the transistor drive. The error amplifier is just a simple op amp, and the transistor drive is a comparator.

If the voltage is too high then the transistor tends more off than on. If the voltage is too low then the transistor tends more on than off. The inductor smooths the output jiggles into a DC voltage.

By only allowing through the required charge, and switching very quickly, the resistive losses are eliminated. If you build the buck regulator from discrete components, it's quite a complicated task. If you buy a buck regulator chip you can just follow the instructions on the datasheet and it will work fine.

An example device is MC33063 which can be used in buck, boost and inverting configurations. It's a £0.40 8 pin part, easily available everywhere despite the chip shortage.

 

Edited By Andy Ash on 10/11/2022 01:45:52

[Martin KyteParticipant @martinkyte99762]

That’s not a bad description of a buck regulator but for driving a string of LEDs if the feedback is taken from a current sensing resistor usually less than 1 ohm then the cct becomes a constant current generator and there is no requirement for any series resistor to the LEDs.

regards Martin

[Michael GilliganParticipant @michaelgilligan61133]

Further to yesterday’s digression … Here’s a short note about the usage of BSS138

**LINK** https://www.elprocus.com/bss138-mosfet/

… and I will also mention that the boards I purchased are now even cheaper mainly because they are listed with free postage.

I am now racking my three remaining neurones to think of uses for them.

MichaelG.

.

[ digression offered in the sprit of Robin’s desire to learn ]

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Andy Ash on 10/11/2022 01:44:22:

I've not read the whole thread, so forgive me if I've repeated something someone else said. Looking at this, it's an ideal application for a "buck regulator". The problem you have is the relatively high power dissipation dropping from 12 to 2ish volts at 270mA.

[…]

.

Mine was the fifth post on the thread, Andy

With that said; your concise description should be very welcome.

MichaelG.

[Robin GrahamParticipant @robingraham42208]

Posted by Andy Ash on 10/11/2022 01:44:22:

I've not read the whole thread, so forgive me if I've repeated something someone else said. Looking at this, it's an ideal application for a "buck regulator". The problem you have is the relatively high power dissipation dropping from 12 to 2ish volts at 270mA.

That's always going to be 2.7 watts in any linear situation. Its not that you can't get a resistor or transistor to handle that, but it's wasteful – especially if you're using a 12V battery.

You can get a buck regulator ready made circuit from ebay, but you can make one too. Broadly it is an oscillator which switches a pass transistor through a diode and an inductor in a PWM fashion. You could use a 555 timer for the oscillator. There is a feedback loop and an error amplifier which are used to control the mark space ratio of the transistor drive. The error amplifier is just a simple op amp, and the transistor drive is a comparator.

If the voltage is too high then the transistor tends more off than on. If the voltage is too low then the transistor tends more on than off. The inductor smooths the output jiggles into a DC voltage.

By only allowing through the required charge, and switching very quickly, the resistive losses are eliminated. If you build the buck regulator from discrete components, it's quite a complicated task. If you buy a buck regulator chip you can just follow the instructions on the datasheet and it will work fine.

An example device is MC33063 which can be used in buck, boost and inverting configurations. It's a £0.40 8 pin part, easily available everywhere despite the chip shortage.

Edited By Andy Ash on 10/11/2022 01:45:52

Thanks Andy. After mucking about with trial circuits and much excerise of my neurones (I have only two remaining by their reckoning, but they may be wrong) it eventually dawned on me that my original idea bought nothing, for exactly the reasons you have given. Before reading your reply I had actually ordered some Buck Converters which I hope will fit the bill. They should arrive tomorrow.

MichaelG – apologies if you were the first to suggest a buck converter and I didn't immediately take that on board – I'd never heard of the things and it went over my head.

Robin.

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Robin Graham on 13/11/2022 01:57:37:

[…]

MichaelG – apologies if you were the first to suggest a buck converter and I didn't immediately take that on board – I'd never heard of the things and it went over my head.

Robin.

.

No problem at all, Robin … it’s been interesting to follow the discussion,

and I’m sure you now have a much fuller understanding of why it’s probably the optimum solution.

MichaelG.

[Robin GrahamParticipant @robingraham42208]

To draw a line under this, the buck converters (£8 for five delivered!) arrived today. I lashed up a chain of 10 'ultra bright' LEDs in wired in parallel. These things are rated 20,000 mcd, 2.5V max voltage, 50mA max current. I put a 2 ohm limiting resistor in series because the converters are rated at maximum 3A output and I didn't want to blow anything up. I assume that the converters are pure voltage sources and it's up to the user to ensure the load doesn't ask for more than 3A at whatever output voltage is dialed in.

It works beautifully – setting 12V on the supply and 3.6V on the output gave an input current of 130mA and output of 310mA. So a loss of about 440mW. That'll go down when I get it properly wired. The LEDs are eye-burningly bright!

Robin.

[John HaineParticipant @johnhaine32865]

Phew!!

______________________________________________________________________________________________________________

[Martin KyteParticipant @martinkyte99762]

Posted by Robin Graham on 13/11/2022 21:38:51:

To draw a line under this, the buck converters (£8 for five delivered!) arrived today. I lashed up a chain of 10 'ultra bright' LEDs in wired in parallel. These things are rated 20,000 mcd, 2.5V max voltage, 50mA max current. I put a 2 ohm limiting resistor in series because the converters are rated at maximum 3A output and I didn't want to blow anything up. I assume that the converters are pure voltage sources and it's up to the user to ensure the load doesn't ask for more than 3A at whatever output voltage is dialed in.

It works beautifully – setting 12V on the supply and 3.6V on the output gave an input current of 130mA and output of 310mA. So a loss of about 440mW. That'll go down when I get it properly wired. The LEDs are eye-burningly bright!

Robin.

So a loss of about 440mW. Plus 200mW for the 2 ohm resistor.

Fro your amusement here is a job I did a few years back. Worklighting on a PCB Drill.

It uses a Texas Instruments LM3404 and all the control is done by the switcher with the green current sense resistor providing the feedback. Volts in preset current out just connect the LEDs. I generally recon to get somewhere north of 85% efficiency.

regards Martin

[Michael GilliganParticipant @michaelgilligan61133]

As a quick postscript to Robin’s thread …

Anyone wishing to ‘play’ with strings of LED lights could do much worse than dash to Poundland.

[assuming that the store in Flint is representative] There are silly-cheap bargains to be had.

MichaelG.

[Robin GrahamParticipant @robingraham42208]

A post postcript – this project rapidly got out of hand. I'd bought some high intensity LED's and after reading replies to my OP got them working with reasonable efficiency. Then I thought it would be nice to have them pulsing on and off in a random sort of way. I remembered that I have an Arduino Uno which I had never used and thought maybe I could do something with that. But maximum current from from the digital pins is a bit limited, so I bought some MOSFETs. No idea what I was doing so tested them:

That gave me some idea of what they do – I'm very much in the shallow end of the pool I know, but I have to reason things out for myself. Slowly these days!

It gave me confidence to to build a circuit which, with C/R timing networks, gives satisfyingly random pulsed lights.

This electronics thing has now become something of an obsession – I can't stick the workshop temperature (at present about 10C), can't afford to heat it at current prices, but I can do this stuff in the house.

My thanks again to those who have patiently given advice – somehow doing this simple thing has opened a floodgate in my mind.

I had to resort to plan B (no pulsed lights) with the stove but I did do a nice toilet throne which released a helium balloon when the chain was pulled. Completely ridiculous, but the audience seemed to like it.

 

Robin

 

Edited By Robin Graham on 01/01/2023 00:04:32

[duncan webster 1Participant @duncanwebster1]

If you want to drive direct off an Arduino you'll do better with logic level mosfets. They manage with lower drive voltage

[Robin GrahamParticipant @robingraham42208]

Posted by duncan webster on 01/01/2023 00:22:48:

If you want to drive direct off an Arduino you'll do better with logic level mosfets. They manage with lower drive voltage

Thanks Duncan – more useful info. I got around it by making a two-stage driver, but shall look at these devices.

Robin.

[Author]

Posts